[chrismorgan]

A few months ago, Cloudflare accidentally turned off 1.1.1.1 (I’m simplifying slightly, most notably DNS-over-HTTPS continued to work). Over the course of five or six minutes, traffic dropped to 10% of normal, and stayed there. Somehow, it took another six minutes before an alert fired, at which point they noticed.

https://news.ycombinator.com/item?id=44578490

You’d think that for such a company they’d notice if global traffic for one of their important services for a given minute had dropped below 50% compared with the last hour, but apparently not.

And that’s Cloudflare, who I would expect better of than most.

[onionisafruit]

At my company we will notice that traffic drops significantly for a minute, but thanks to reporting latency we don’t get alerted until a few minutes later. In our business and at our scale, that latency is fine because we aren’t a vital internet service.

edit: I should have followed your link before commenting, because this sentiment is well covered there.